Before You Buy Anything
Everything in this section reduces to ten questions, sent before anybody installs anything.
The ten
Purpose
One. What decision will be made differently once this exists? If the answer is "better visibility", ask again about the visibility.
Two. What are we switching off? Adoption depends on something being replaced rather than added.
Money
Three. Total annual cost at our headcount, on the tier that covers our requirements, including anything billed separately.
Four. What triggers the next tier and what does it cost then? Administrative controls commonly sit one level up.
Five. Can the price rise at renewal, and is any increase capped?
Data
Six. What is captured by default, and can anything we do not want be removed rather than hidden? This is where an administrative tool becomes a monitoring one.
Seven. Where is data processed, and who else sees it? For anything holding employee data this matters more than for most.
Eight. What can an employee see of their own data? The difference between 77% acceptance and 54% considering leaving.
Leaving
Nine. What exactly exports, in what format, and what does not?
Ten. What is the non-renewal notice window, in days, and in what form must notice be given?
Why the order matters
Purpose first, because a clear answer to question one eliminates more candidates than the other nine combined — and frequently eliminates the purchase.
Money second, because it is comparable across vendors and settles the shortlist.
Data and leaving last, because they are the ones that cost you later and nobody asks them under time pressure.
Sent before the trial, not after. Two weeks with a product creates commitment, and commitment is the wrong state for reading a contract.
Reading the answers
Speed. Two days for ten routine questions says something about the organisation behind the product.
Specificity. "Standard formats" is not an answer to question nine. "CSV and JSON, self-service, everything except attachment binaries" is.
And what is skipped. An answer omitting six and nine has omitted the two that cost most.
What to do with them
Put the important ones in the agreement. Export scope, the renewal cap, what is captured by default.
File the rest with the contract, where whoever handles the renewal in three years will find them.
Why ten and not thirty
Because thirty does not get sent.
These are the failures that are expensive and invisible — the ones a trial will never reveal, because a trial shows you the interface and tells you nothing about renewal, export or default capture.
The short version
- Ten questions in one email before the trial: two on purpose, three on money, three on data, two on leaving
- Question one eliminates more candidates than the other nine combined, and frequently eliminates the purchase
- Ask what is captured by default and whether unwanted capture can be removed rather than hidden
- Ask what an employee can see of their own data, which is the difference between acceptance and resignation risk
- Send before the trial, because two weeks with a product creates commitment and commitment is the wrong state for reading a contract
- Judge the answers on speed, specificity and what was skipped, then put export scope, the renewal cap and default capture into the agreement
For additional context on this topic, see Nintendo.